计算机辅助设计与图形学学报
計算機輔助設計與圖形學學報
계산궤보조설계여도형학학보
JOURNAL OF COMPUTER-AIDED DESIGN & COMPUTER GRAPHICS
2014年
11期
2016-2022
,共7页
赵凡%赵建%张葆%尹传历
趙凡%趙建%張葆%尹傳歷
조범%조건%장보%윤전력
图像增强%直方图规定化%子层分割
圖像增彊%直方圖規定化%子層分割
도상증강%직방도규정화%자층분할
image enhancement%histogram specification%sub-layer segmentation
为了更好地辨别低照度图像中的目标,提高图像的对比度并增强纹理细节信息,采用子层分割规定化算法对图像进行增强处理,提出一种结合子层分割和自适应函数引导的规定化图像增强算法。首先根据原始图像直方图的特点对直方图进行子层分割,之后通过对分割后的子层进行灰度区间拉伸来提高图像的亮度,以增强图像的对比度;为了增强图像中的细节信息,在每个子层求取自适应函数来对每个子层做规定化处理;最后对子层直方图合并,得到增强后图像。实验结果证明,经过增强后图像的灰度平均梯度值为原始图像的3~4倍,信息熵也明显增大;该算法在增强图像细节、提高图像的对比度上具有优越性。
為瞭更好地辨彆低照度圖像中的目標,提高圖像的對比度併增彊紋理細節信息,採用子層分割規定化算法對圖像進行增彊處理,提齣一種結閤子層分割和自適應函數引導的規定化圖像增彊算法。首先根據原始圖像直方圖的特點對直方圖進行子層分割,之後通過對分割後的子層進行灰度區間拉伸來提高圖像的亮度,以增彊圖像的對比度;為瞭增彊圖像中的細節信息,在每箇子層求取自適應函數來對每箇子層做規定化處理;最後對子層直方圖閤併,得到增彊後圖像。實驗結果證明,經過增彊後圖像的灰度平均梯度值為原始圖像的3~4倍,信息熵也明顯增大;該算法在增彊圖像細節、提高圖像的對比度上具有優越性。
위료경호지변별저조도도상중적목표,제고도상적대비도병증강문리세절신식,채용자층분할규정화산법대도상진행증강처리,제출일충결합자층분할화자괄응함수인도적규정화도상증강산법。수선근거원시도상직방도적특점대직방도진행자층분할,지후통과대분할후적자층진행회도구간랍신래제고도상적량도,이증강도상적대비도;위료증강도상중적세절신식,재매개자층구취자괄응함수래대매개자층주규정화처리;최후대자층직방도합병,득도증강후도상。실험결과증명,경과증강후도상적회도평균제도치위원시도상적3~4배,신식적야명현증대;해산법재증강도상세절、제고도상적대비도상구유우월성。
In order to distinguish the target of the illumination image better 、increase the contract and enhance the weak information ,by using sub‐layer segmentation and histogram specification to enhance image , we propose the image enhancement combined sub‐layer segmentation with histogram specification induced by adaptive function algorithm .Firstly ,according to the features of the original image ,its histogram will be divided into sub‐layer .Then ,stretching gradation interval of the sub‐layer to increase the image brightness ,the image contrast can be enhanced . To enhance the image details better ,we use the required adaptation functions to do the histogram specification in each sub‐layer .Finally ,the sub‐layer histogram will be merged to get the enhanced image .Experimental results show that the gray mean gradient value of the enhanced image was 3‐4 times of the original image ,and entropy was increased obviously .The algorithm can enhance image detail and improve image contrast with superiority .